China is preparing to approve limited imports of Nvidia’s (NVDA) H200 artificial intelligence chips as soon as this quarter.

The move would partially restore access to the Chinese market for the U.S. chipmaker after several years of export restrictions tied to national security concerns.

People familiar with the matter said approval would apply only to selected commercial uses. The H200 would remain prohibited for military applications, sensitive government agencies, critical infrastructure, and most state-owned enterprises.

Some limited exceptions could apply, according to the report.

The H200 is an older-generation chip based on Nvidia’s Hopper architecture. It is used to train large artificial intelligence models.

The Trump administration previously allowed exports of the H200 to China. Newer processors, including the Blackwell platform and future Rubin chips, remain restricted.

Last week, several major Chinese technology companies have expressed interest in the H200.

Those companies include Alibaba Group and ByteDance, which are seeking to expand AI capabilities as competition with Western firms intensifies.

Nvidia plans to ship between 5,000 and 10,000 chip modules to China in the coming months.

That volume would be equivalent to roughly 40,000 to 80,000 individual H200 chips.

A reopening of the market, even on a limited basis, would mark a significant development for Nvidia’s China business.

Nvidia Chief Executive Jensen Huang has said previously that China’s AI chip market could reach $50 billion within a few years.

Executives said demand from Chinese customers for the H200 remains strong.

They added that Nvidia has not held direct discussions with Beijing regarding approval timing.

The company has submitted license applications to U.S. authorities, and final export approval details are still being finalized.

In New York, Nvidia shares were marginally higher in premarket trading following the report.